Paradoxically, children form strong attachments to their abusive caregiver, and these children are at high risk for psychiatric disorders both during childhood and as adults. Although this is an important ethical /societal and mental health problem, how this paradoxical attachment can occur remains a mystery. Specifically the neurological underpinnings of this paradoxical attachment are not known;however there are substantial data showing that the amygdala has an important regulatory role in fear/aversion learning both in adults and infants. We have modeled the neurobiology of abusive attachment using neonatal rats and fear conditioning (odor-shock pairings). We found odor-aversive shock pairings produces an odor preference during a sensitive period of development in neonatal rats but an aversion in older pups and adults. The naturally occurring elevation in corticosterone that occurs around 10 days of age in the rat pup appears to bring the amygdala "on-line" and switch the infant to the adult response of avoidance from odor-shock conditioning. Specifically, we can either accelerate or retard the end of the sensitive period simply by respectively lowering or raising the pup's endogenous CORT levels either systemically or selectively within the amygdala. Because these learned preferences to aversive stimuli are long lasting they likely engage changes in genes and proteins within the amygdala. We use here microarrays to assess gene expression during and after the "sensitive" period to understand what the mechanisms are within the amygdala. We determine: First, what are the changes in gene expression induced by noxious input when preferences or aversions are learned? This provides a detailed description of how the amygdala responds differently before and after the switch from preference to aversion. Second, what changes in gene expression occur in the amygdala as a result of manipulation of CORT levels that also alter the sensitive period? Third, what are the long term changes in aversion or preference learning that persist into adulthood and modify levels of engagement in other motivated behaviors? The combined use of age-specific learning and CORT (altering the age that preference learning switches to aversion learning), provides a powerful tool to assess amygdala candidate genes in an age dependent and independent paradigm. Our preliminary studies show contrasting changes in dopamine markers during and after the sensitive period and follow-up studies implicate further the role of dopamine in the learned preferences to aversive stimuli.